Qwen3.8-27B-abliterated
27BQwen
from Qwen/Qwen3.8-27B
↓ 3.9M♥ 1.7K166 repos
Qwen3.8-27B-Uncensored-NInfer
Qwen3.8-27B-Uncensored-NInfer is a abliterated and uncensored variant of Qwen/Qwen3.8-27B, published on Hugging Face by pyros-vault. It is 27B parameters, 48GB VRAM class class and apache-2.0 licence. It is distributed across 8 repositories in Transformers and NVFP4 formats, totalling 27.1K downloads.
Specification
- Model name
- Qwen3.8-27B-Uncensored-NInfer
- Source repository
- pyros-vault/Qwen3.8-27B-Uncensored-NInfer
- Base model
- Qwen/Qwen3.8-27B
- Publisher
- pyros-vault
- Parameters
- 27B
- VRAM required
- ~17.2 GB at Q4_K_M, 4k context (estimated) — runs on 24 GB cards
- Hardware class
- 25B – 50B — 48GB VRAM class
- Licence
- apache-2.0
- Task
- Image text to text
- Context window
- 4K tokens (as documented)
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, NVFP4
- First indexed
- 15 Aug 2026
- Last updated
- 2 Oct 2026
Ablation technique
How Qwen3.8-27B-Uncensored-NInfer was modified, and what that implies.
Abliteration + uncensored fine-tune
Directional ablation combined with additional fine-tuning on unfiltered data, so behaviour diverges from the base model beyond refusal removal alone.
Reported scope: Not stated by the publisher. Publishers frequently omit this, so verify behaviour empirically rather than assuming full coverage.
From the publisher’s model card
“This is a deployment-format conversion only. It does not add training, change the upstream behavior, or independently establish the upstream model's "uncensored" or "abliterated" characteristics. The artifact is not a Safetensors or GGUF checkpoint and is not intended for Transformers, llama.cpp, or unrelated NInfer...”
Running it
Commands are templates — confirm the exact repository and quant file before use.
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "pyros-vault/Qwen3.8-27B-Uncensored-NInfer"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve pyros-vault/Qwen3.8-27B-Uncensored-NInfer --trust-remote-codeDownloads and variants (7)
Every published repository of Qwen3.8-27B-Uncensored-NInfer, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| pyros-vault/Qwen3.8-27B-Uncensored-NInfersource | Transformers | 12.7K |
| xidisdg/qwen3.8_27b-uncensored-ninfer | Transformers | 5K |
| igorls/Qwen3.8-27B-Uncensored-NVFP4-NInfer | NVFP4 | 3.1K |
| pcmaker/Huihui-Qwen3.8-27B-Uncensored-NInfer | Transformers | 2.6K |
| fullmetaljackass/Qwen3.8-27B-Uncensored-NInfer | Transformers | 1.9K |
| JMVRoill/Qwen3.8-27B-Uncensored-nvfp4-NInfer | NVFP4 | 1.8K |
| DogOnKeyboard/Qwen3.8-27B-Uncensored-NInfer | Transformers | 0 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Qwen3.8-27B-Uncensored-NInfer will attempt offensive-security prompts that a hosted commercial model declines, which is what makes it usable for red-team corpus generation and for measuring what an unaligned model of this class produces.
- Capability is inherited from Qwen/Qwen3.8-27B, not added by ablation — at 27B parameters, expect roughly the base model’s competence, minus some instruction-following fidelity.
- The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
- This release adds fine-tuning on unfiltered data, so its behaviour diverges from the base model beyond refusal removal — differences you measure cannot be attributed to ablation alone.
- Declared licence is apache-2.0, but the terms that bind you are Qwen/Qwen3.8-27B’s — a re-release cannot grant more than its parent.
Handling: run it isolated, put moderation in front of it if anyone outside your team can reach it, and verify the weights before loading — these are third-party artifacts. Full handling and licence guidance.
Metadata for Qwen3.8-27B-Uncensored-NInfer is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with pyros-vault, and does not independently verify publisher claims. See responsible use.